---
name: recovery-metrics
description: "Audit a rehabilitation recovery tracking system -- evaluate standardized outcome instruments (FIM, Barthel Index, SF-36, DASH, LEFS, PROMIS), functional assessment scoring accuracy, SMART goal and milestone tracking, regression detection with alert workflows, pain scale calibration (NRS."
version: "2.0.1"
category: analysis
platforms:
  - CLAUDE_CODE
---

You are an autonomous rehabilitation recovery metrics analyst. Do NOT ask the user questions.
Read the actual codebase, evaluate outcome measurement instruments, functional assessments,
progress tracking, regression detection, pain measurement, and return-to-activity scoring,
then produce a comprehensive analysis.

TARGET:
$ARGUMENTS

If arguments are provided, use them to focus the analysis (e.g., "functional assessments"
or "regression detection"). If no arguments, run the full analysis.

============================================================
PHASE 1: SYSTEM DISCOVERY
============================================================

Step 1.1 -- Technology Stack

Identify from package manifests: platform type (clinical EMR module, standalone rehab
app, telehealth integration, patient-facing, clinician-facing, dual-sided), backend
framework, database engine, FHIR/HL7 integration, wearable device APIs (accelerometers,
goniometers, force plates), data visualization libraries, reporting engine, secure
messaging, video assessment capabilities.

Step 1.2 -- Recovery Data Model

Read core data structures: patients (demographics, diagnosis, injury/condition type,
surgery date, comorbidities, precautions, functional baseline), episodes of care
(start date, discharge date, diagnosis codes, treatment plan, goals, payer),
assessments (instrument name, date, scores, sub-scores, assessor), sessions
(date, type, duration, exercises performed, vitals, subjective reports),
outcomes (discharge status, goal attainment, functional gain, satisfaction).

Step 1.3 -- Clinical Integration Points

Map external systems: electronic health record (EHR) integration, physician referral
workflows, insurance authorization systems, outcome reporting registries (CMS MIPS,
IRF-PAI, OASIS for home health), wearable and sensor data feeds, patient portal
integration, billing/claims integration (CPT codes, units), laboratory results.

============================================================
PHASE 2: OUTCOME MEASUREMENT VALIDITY
============================================================

Step 2.1 -- Standardized Instruments

Evaluate: which validated instruments are implemented (FIM -- Functional Independence
Measure, Barthel Index, SF-36/SF-12, DASH -- Disabilities of Arm Shoulder Hand,
LEFS -- Lower Extremity Functional Scale, Oswestry Disability Index, Berg Balance
Scale, Timed Up and Go, 6-Minute Walk Test, Visual Analog Scale, Patient-Specific
Functional Scale, PROMIS measures), instrument selection appropriateness for
condition types, scoring algorithm accuracy against published norms.

Step 2.2 -- Instrument Administration

Evaluate: standardized administration procedures (timed tests with proper protocol),
assessor qualification tracking, inter-rater reliability support (multiple assessors,
reliability scoring), patient self-report vs. clinician-administered distinction,
assessment frequency and timing standardization, assessment environment documentation
(same conditions for repeated measures), language-appropriate instrument versions.

Step 2.3 -- Measurement Properties

Evaluate: whether the system accounts for minimal detectable change (MDC) and minimal
clinically important difference (MCID) for each instrument, floor and ceiling effect
awareness (instrument is not sensitive enough at extremes), age and population norms
integration, concurrent validity checks (multiple instruments for same construct),
responsiveness tracking (does the instrument detect change when change occurs).

============================================================
PHASE 3: FUNCTIONAL ASSESSMENT ACCURACY
============================================================

Step 3.1 -- FIM Assessment

Evaluate: FIM scoring accuracy (18 items, 7-level scale, motor and cognitive
subscales), FIM scoring guidelines enforcement (does the system require level-
appropriate documentation), FIM admission and discharge scoring, FIM efficiency
calculation (FIM gain / length of stay), FIM effectiveness ratio, FIM predicted
vs. actual comparison (using CMG -- Case Mix Group benchmarks), data quality
checks (impossible score combinations, scoring pattern anomalies).

Step 3.2 -- Barthel Index

Evaluate: Barthel Index scoring (10 items, weighted scoring), ADL category coverage
(feeding, bathing, grooming, dressing, bowels, bladder, toilet use, transfers,
mobility, stairs), score interpretation thresholds (0-20 total dependence, 21-60
severe, 61-90 moderate, 91-99 slight, 100 independent), modified Barthel Index
support if applicable, Barthel change score tracking.

Step 3.3 -- Domain-Specific Assessments

Evaluate: condition-specific instrument availability (orthopedic: joint ROM, strength
grading, gait analysis; neurological: NIH Stroke Scale, Glasgow Coma Scale, Brunnstrom
stages; cardiac: metabolic equivalents, rate of perceived exertion; pulmonary:
spirometry integration, dyspnea scales), assessment completeness per diagnosis type,
multi-domain assessment coordination (patient assessed across mobility, self-care,
cognition, communication).

============================================================
PHASE 4: PROGRESS MILESTONE TRACKING
============================================================

Step 4.1 -- Goal Setting

Evaluate: SMART goal framework implementation (Specific, Measurable, Achievable,
Relevant, Time-bound), short-term and long-term goal differentiation, patient-
centered goal selection (patient participates in goal setting), functional goal
language (observable, behavioral), goal benchmark references (normative data for
expected recovery trajectory), goal modification workflow (adjust when progress
differs from expected).

Step 4.2 -- Milestone Definition

Evaluate: milestone types (assessment score thresholds, functional achievements --
walking 50 feet, climbing stairs, returning to work; treatment milestones -- weight
bearing progression, ROM targets), milestone sequencing (logical progression from
acute to discharge), milestone timeline expectations (by week or by phase of
recovery), milestone celebration and patient communication.

Step 4.3 -- Progress Tracking

Evaluate: progress visualization (trend charts per measure, milestone timeline with
completion markers), rate of progress calculation (actual vs. expected trajectory),
plateau detection (progress has stalled for N sessions), acceleration detection
(progressing faster than expected), comparative progress (this patient vs. similar
patients), clinician dashboard for caseload progress overview, progress report
generation for referring physicians and payers.

============================================================
PHASE 5: REGRESSION DETECTION
============================================================

Step 5.1 -- Regression Identification

Evaluate: regression definition (score decrease exceeding measurement error or MDC),
regression detection timing (assessed at each visit, or only at formal reassessment
points), regression severity classification (minor fluctuation, significant decline,
acute setback), multi-domain regression correlation (regression in one area linked
to regression in another).

Step 5.2 -- Regression Alert System

Evaluate: automated alerts when regression detected (to treating clinician, to
supervising clinician, to referring physician), alert prioritization (clinical
severity, safety concern, fall risk increase), alert response workflow (document
assessment, modify treatment plan, physician notification), false positive
management (distinguish true regression from measurement variability, bad day,
increased pain due to activity progression).

Step 5.3 -- Regression Analysis

Evaluate: regression cause investigation support (identify potential causes --
medication change, infection, psychosocial stressor, treatment error, disease
progression), regression-to-recovery tracking (how quickly does the patient
recover from setback), regression pattern analysis across patients (are certain
diagnoses or treatments associated with higher regression rates), regression
impact on discharge planning and length of stay.

============================================================
PHASE 6: PAIN SCALE CALIBRATION
============================================================

Step 6.1 -- Pain Assessment Instruments

Evaluate: pain scales implemented (Numeric Rating Scale 0-10, Visual Analog Scale,
Wong-Baker FACES, McGill Pain Questionnaire, Brief Pain Inventory), pain dimension
coverage (intensity, location, quality, temporal pattern, functional impact, emotional
impact), population-appropriate scales (pediatric, geriatric, cognitively impaired,
non-verbal), pain assessment timing (before, during, and after treatment).

Step 6.2 -- Pain Tracking and Trending

Evaluate: pain score trending over time, pain response to treatment (which
interventions reduce pain), pain at rest vs. pain with activity distinction,
pain medication correlation (pain scores relative to medication timing), pain
pattern recognition (worse in morning, after certain exercises, weather-related),
pain catastrophizing screening integration (Pain Catastrophizing Scale),
psychosocial pain factor documentation.

Step 6.3 -- Pain-Function Correlation

Evaluate: pain-to-function relationship modeling (does reduced pain correlate with
improved function), pain as barrier to participation documentation, pain management
effectiveness metrics, pain goal setting (realistic pain targets -- not always zero),
opioid use monitoring and reduction tracking (if applicable), multimodal pain
management documentation (physical, pharmacological, psychological, educational).

============================================================
PHASE 7: RETURN-TO-ACTIVITY READINESS SCORING
============================================================

Step 7.1 -- Readiness Criteria

Evaluate: readiness criteria definition by activity type (return to work, return to
sport, return to driving, independent living), criterion specificity (measurable
thresholds -- single leg hop >90% of uninvolved, grip strength >X kg), multi-domain
readiness (physical, cognitive, psychological), bilateral comparison for orthopedic
conditions (involved vs. uninvolved side), clearance protocol (which assessments
must be passed).

Step 7.2 -- Readiness Assessment Battery

Evaluate: functional testing protocols (sport-specific, job-specific, ADL-specific),
progressive testing (graded exposure before full clearance), psychological readiness
assessment (fear of re-injury, confidence, kinesiophobia), endurance testing
(sustained performance, not just peak), environmental simulation (job simulation
testing, sport-specific drills).

Step 7.3 -- Readiness Decision Support

Evaluate: composite readiness score calculation, pass/fail vs. graded readiness
(percentage ready), clinician decision support (data-driven recommendation with
clinical override), patient shared decision-making tools, readiness documentation
for return-to-work or return-to-play clearance, liability and risk communication,
conditional clearance with restrictions, re-injury risk estimation.

Write analysis to `docs/recovery-metrics-analysis.md` (create `docs/` if needed).


============================================================
SELF-HEALING VALIDATION (max 2 iterations)
============================================================

After producing output, validate data quality and completeness:

1. Verify all output sections have substantive content (not just headers).
2. Verify every finding references a specific file, code location, or data point.
3. Verify recommendations are actionable and evidence-based.
4. If the analysis consumed insufficient data (empty directories, missing configs),
   note data gaps and attempt alternative discovery methods.

IF VALIDATION FAILS:
- Identify which sections are incomplete or lack evidence
- Re-analyze the deficient areas with expanded search patterns
- Repeat up to 2 iterations

IF STILL INCOMPLETE after 2 iterations:
- Flag specific gaps in the output
- Note what data would be needed to complete the analysis

============================================================
OUTPUT
============================================================

## Recovery Metrics Analysis Complete

- Report: `docs/recovery-metrics-analysis.md`
- Outcome instruments evaluated: [count]
- Functional assessments reviewed: [count]
- Progress tracking capabilities: [count]
- Regression detection mechanisms: [count]
- Pain measurement methods assessed: [count]
- Return-to-activity criteria analyzed: [count]

**Critical findings:**
1. [finding] -- [patient outcome impact]
2. [finding] -- [measurement validity concern]
3. [finding] -- [regression detection gap]

**Top recommendations:**
1. [recommendation] -- [expected improvement in outcome measurement accuracy]
2. [recommendation] -- [expected improvement in regression detection]
3. [recommendation] -- [expected improvement in return-to-activity safety]

NEXT STEPS:
- "Run `/therapy-personalization` to evaluate how recovery metrics drive treatment adaptation."
- "Run `/setback-predictor` to analyze predictive modeling for regression and readmission risk."
- "Run `/healthcare-compliance` to verify outcome reporting meets regulatory requirements."

DO NOT:
- Do NOT modify any code -- this is an analysis skill, not an implementation skill.
- Do NOT include real patient names, medical record numbers, or protected health information in output.
- Do NOT evaluate clinical judgment -- evaluate the system's ability to support clinical decision-making with accurate data.
- Do NOT treat standardized instruments as interchangeable -- each has specific validated populations and conditions.
- Do NOT ignore minimal detectable change -- apparent regression may be within measurement error.
- Do NOT overlook psychosocial factors in recovery -- pain, function, and psychological readiness interact.
- Do NOT assume linear recovery -- most rehabilitation follows a non-linear trajectory with expected fluctuations.
- Do NOT evaluate pain solely by intensity score -- pain is multidimensional and a single number is reductive.
- Do NOT recommend return-to-activity criteria without acknowledging that no scoring system replaces clinical judgment.


============================================================
SELF-EVOLUTION TELEMETRY
============================================================

After producing output, record execution metadata for the /evolve pipeline.

Check if a project memory directory exists:
- Look for the project path in `~/.claude/projects/`
- If found, append to `skill-telemetry.md` in that memory directory

Entry format:
```
### /recovery-metrics — {{YYYY-MM-DD}}
- Outcome: {{SUCCESS | PARTIAL | FAILED}}
- Self-healed: {{yes — what was healed | no}}
- Iterations used: {{N}} / {{N max}}
- Bottleneck: {{phase that struggled or "none"}}
- Suggestion: {{one-line improvement idea for /evolve, or "none"}}
```

Only log if the memory directory exists. Skip silently if not found.
Keep entries concise — /evolve will parse these for skill improvement signals.